Related stories
One that is close to my heart, as a long-time and still keen Skype user, is the matter of QoS (Quality of Service) including reliability and voice clarity. In this case, I'm hoping to get feedback from Skype themselves, not the consumer support desk but from a Skype executive/architect or two with in-depth knowledge.
I first started using Skype back in its Beta testing days during late 2003 and early 2004. Even back then, given a reasonably fast broadband connection (such as the Optus Cable that I've been using all along), the voice quality for PC-to-PC calls was impressive. Surprisingly, over slow dial-up connections the quality was often also quite tolerable, much better than what I recall of the Voice over Internet Protocol (VoIP) attempts in the mid-to-late-1990s, when the likes of companies such as Net2Phone came and went.
For PC-based calls, the Skype peer-to-peer (P2P) architecture generally works quite well -- most of the time. The spectacular mid-August 2007 failure just proves that there's always a distinction to be made between architecture and implementation: see Skype users don't disconnect your PSTN service just yet and "Perfect storm" sank Skype. (Presumably they've now tweaked the network self-healing parameters to cope with future storms of this kind.)
You can gain some insights into the architecture in the September 2004 paper by Salman A. Baset and Henning Schulzrinne of the Department of Computer Science at Columbia University, New York: " An Analysis of the Skype Peer-to-Peer Internet Telephony Protocol" and from the official Skype Guide for Network Administrators.
Article Link at http://www.itwire.com/content/view/14283/1127/
